Added error or warning facilitation for unscheduled algs. #1734
Conversation
|
/deploy |
Adding nodes resource for error message info, WIP.
|
/deploy |
Added ability for pipeline-driver to detect warnings vs errors, upon unScheduling
|
/deploy |
|
/deploy |
|
/deploy |
|
Error or warning distinction, format fixes, node graph object updates concerning batch use-case. |
1 similar comment
|
Error or warning distinction, format fixes, node graph object updates concerning batch use-case. |
There was a problem hiding this comment.
Reviewed all commit messages.
Reviewable status: 0 of 7 files reviewed, 4 unresolved discussions (waiting on @RonShvarz)
core/pipeline-driver/lib/datastore/graph-store.js line 128 at r5 (raw file):
_handleBatch(n) { const isAnyBatchFailed = n.batch.any(b => b.status === 'FailedScheduling');
anyFailedScheduling
Code quote:
isAnyBatchFailed core/pipeline-driver/lib/state/state-manager.js line 121 at r5 (raw file):
if (resources && resources[0] && resources[0].unScheduledAlgorithms) { const algorithms = { ...resources[0].unScheduledAlgorithms, ...resources[0].ignoredUnscheduledAlgorithms }; const nodesFromEtcd = resources[0].nodes;
clusterNodes
Code quote:
nodesFromEtcdcore/pipeline-driver/lib/tasks/task-runner.js line 123 at r5 (raw file):
n.status = event.reason; n.warnings = n.warnings || []; const { resourceMessage, isError } = this._nodeResourceMessageBuilder(event, nodesFromEtcd);
buildNodeRe....
Code quote:
_nodeResourceMessageBuildercore/pipeline-driver/lib/tasks/task-runner.js line 156 at r5 (raw file):
else { resourceMessage += `${unScheduledAlg.complexResourceDescriptor.numUnmatchedNodesBySelector} nodes don't match node selector: '${selectors}',\n`; ({ resourceMessage, isError } = this._specificNodesResourceMessageBuilder(unScheduledAlg, resourceMessage, isError, nodesFromEtcd));
buildSpecif....
Code quote:
specificNodesResourceMessageBuilde
There was a problem hiding this comment.
Reviewable status: 0 of 7 files reviewed, 5 unresolved discussions (waiting on @RonShvarz)
core/pipeline-driver/lib/tasks/task-runner.js line 191 at r5 (raw file):
resourceMessage += `${k} ,\n`; const currentValue = resourceIdentifier.get(k); resourceIdentifier.set(k, currentValue + 1);
BreachCountPerResource
Code quote:
resourceIdentifierCR changes and fixes, Robust over-capacity warning, Unit tests overhaul
|
/deploy |
|
/deploy |
…hen it is not the first in the running order
|
/deploy |
There was a problem hiding this comment.
Reviewable status: 0 of 7 files reviewed, 5 unresolved discussions (waiting on @golanha)
core/pipeline-driver/lib/datastore/graph-store.js line 128 at r5 (raw file):
Previously, golanha (Golan Hallel) wrote…
anyFailedScheduling
Done.
core/pipeline-driver/lib/state/state-manager.js line 121 at r5 (raw file):
Previously, golanha (Golan Hallel) wrote…
clusterNodes
Done.
core/pipeline-driver/lib/tasks/task-runner.js line 123 at r5 (raw file):
Previously, golanha (Golan Hallel) wrote…
buildNodeRe....
Done.
core/pipeline-driver/lib/tasks/task-runner.js line 156 at r5 (raw file):
Previously, golanha (Golan Hallel) wrote…
buildSpecif....
Done.
core/pipeline-driver/lib/tasks/task-runner.js line 191 at r5 (raw file):
Previously, golanha (Golan Hallel) wrote…
BreachCountPerResource
Done.
There was a problem hiding this comment.
Reviewed 2 of 6 files at r3, 3 of 4 files at r6, 1 of 1 files at r7, 1 of 1 files at r8, 1 of 1 files at r9, all commit messages.
Reviewable status:complete! all files reviewed, all discussions resolved (waiting on @RonShvarz)
* Added error or warning facilitation for unscheduled algs. * Fixed unit tests, Adding nodes resource for error message info, WIP. * Edited capital letters in task-executer, Added ability for pipeline-driver to detect warnings vs errors, upon unScheduling * Added warnings and errors in case a node has batches in it. * clearing the node object if there are no failures to schedule * Fix wrong function usage @_handleBatch, CR changes and fixes, Robust over-capacity warning, Unit tests overhaul * Round up values of missing resources to 2 decimal points. * handle use-case of node stuck in prescheduling with failing batches when it is not the first in the running order * fixed formatting for resources .... bump version [skip ci]
* Added error or warning facilitation for unscheduled algs. * Fixed unit tests, Adding nodes resource for error message info, WIP. * Edited capital letters in task-executer, Added ability for pipeline-driver to detect warnings vs errors, upon unScheduling * Added warnings and errors in case a node has batches in it. * clearing the node object if there are no failures to schedule * Fix wrong function usage @_handleBatch, CR changes and fixes, Robust over-capacity warning, Unit tests overhaul * Round up values of missing resources to 2 decimal points. * handle use-case of node stuck in prescheduling with failing batches when it is not the first in the running order * fixed formatting for resources .... bump version [skip ci]
#1638
This change is